Our goal is to use theoretical models, gene expression arrays, molecular genetic markers (SNPs and microsatellites) and molecular evolutionary theory to better understand how gene interactions (epistasis) produce complex phenotypes. As our experimental system, we will use recently collected populations of the red flour beetle, Tribolium castaneum, which exhibit characteristics of incipient speciation in the form of inter-population hybrids with impaired behavior, morphological deformities, and reduced fertility and viability. Our specific aims include (1) confirming the functional identity of a pair of genes from different populations that cause a total reduction of hybrid fitness when brought together in a common genetic background;(2) characterizing the molecular evolution of these genes through their patterns of expression and sequence diversity within and among species;(3) identifying, confirming and characterizing a different set of interacting genes which impair the motor behavior of inter-population hybrids;(4) characterizing the differences in gene expression between inter-population hybrids and their parents using Nimblegen[unreadable] microarrays;and, (5) extending and testing two-locus population genetic theory to a new model of speciation based on interactions between a pair of genes, one with maternal and one with zygotic expression. Our extensive preliminary studies, crossing pairs of populations from North and South America, Africa and Asia, have revealed that dominance, epistasis, maternal genetic effects, and genotype-by-environment interaction cause morphological and behavioral abnormalities and reduced viability and fertility in inter-population hybrids. For the most extreme case of hybrid fitness reduction, we have shown, through repeated back-crossing to the genome strain, GA-2, and microsatellite marker mapping, that a specific pair of genes interacts to reduce hybrid fitness. We have also shown that inter-population behavioral abnormalities occur in matings and backcrosses between a different pair of populations; that they are not caused one of the genes responsible for the extreme reduction of hybrid fitness (above);and that they involve the breakdown of linked gene combinations. By continuing our back-crosses to the genomic strain, we can isolate the gene combinations causing the behavioral abnormalities in a common genetic background, confirm their functional identity, and characterize their evolution within and among species.